I’m having an issue, that my 8TB node has a pretty high I/O wait during startup, which holds on for up to several days on my Odroid HC2. I can remember, that there was an option, to disable the scan on startup. I’m using docker, unfortunately I can’t find the thread in the forum, where the command was mentioned, which I can add to prevent the scan on startup. I hope, you can help me out. Thanks in advance
sorry for the hijack in thread, i face a similar issye on my windows node. for varius reason, my windows get a bsod every 3-4 days, at the moment cant recornige the cause of bsod.
in that case, after every startup, my 8tb hdd (cmr) have high I/O for almost 48hours, so many times, even cant complete the scan? and another bsod coming.
is there any similar option
–storage2.piece-scan-on-startup=false
for windows node? in config there isn’t this option. can i add manualy?
also, is this best practice? or maybe face another issues with node?
my node is almost 30month old
It makes no sense to continue tweaking storagenode unless you find and address the bsod. Filewalker is actually beneficial for you, it primes the caches, among other things, so your node responds faster and you earn more money.
The reason for the bsod is printed at the bottom of the report. Furthermore, you can open the memory.dmp from system folder in WinDbg (from store) and run !analyze -v to get some diagnostic, at least you’ll know which driver caused it.
Separately, I’d recommend running a memory test (see memtest86) and chkdsk c: /f. Often bad ram or filesystem errors can manifest by spurious panics like these, and the latter tend to get worse after each abrupt reboot.
I changed all my config.yaml to the command you suggested, but unfortunately, it’s not working. after restard, I still have a high I/O wait. Did I made a mistake by any chance?
@deathlessdd@donald.m.motsinger I restarted the system, removed the containers and also did the docker run command in addition, at first there was pretty high I/O wait, thats why I thouhgt it’s not working, but now it seem to be gone, faster than normal, so I think it might have worked.